AN ABSTRACT OF THE THESIS OF
RONALD KEITH REED
(Name)
in
OCEANOGRAPJ-IY
OCEANOGRAPHY
(Major)
Abstract approved:
for the
presented on
MASTER OF SCIENCE
(Degree)
A
;;
:; /L
(Date)
Redacted for privacy
Dr. William P. Elliott
The objective of this study was to ascertain the magnitude and
distribution of rainfall over coastal waters of the northwestern United
States and to compare values with those at nearby land stations.
Precipitation was measured with gages at Totem, rainfall amounts
were assessed from weather reports at lightships off the coast, and
preci.pitation frequencies
precipitation
frequencies at
at lightships
lightships and land stations were examexammed.
Results from the three methods were quite consistent; precipitati.on at sea was only about one-third that at coastal land stations.
These values are appreciably less than previous estimates of oceanic
rainfall in this area, and they support the view that a significant
horizontal gradient of precipitation may exist between the coast and
typically occu.rs
occurs both at
open sea. Rainfall typically
at sea and ashore on the
same day, but it rairis
rains fewer hours at sea. The relative amount of
rain at sea varies with the type of atmospheric system, and rainfall
at the coast appears to be intensified by frictional processes.
Estimates of evaporation minus precipitation are less negative
than earlier ones; consideration of their relation to surface salinity
leads to distributions that are in good agreement with oceanographic
knowledge. The newer values suggest that in this region the heat
gain by the atmosphere may be less (but moisture entrainment may
be greater) than was thought.
Rainfall over Coastal Waters of
the Pacific Northwest
by
Ronald Keith Reed
A THESIS
submitted to
Oregon State University
in partial fulfillment of
the requirements for the
degree of
Master of Science
June 1973
APPROVED:
Redacted for privacy
Redacted
for privacy
Oceanography
Associate Professor of O
in Redacted for privacy
Redacted for privacy
Dean g'Scbo-Y
iooY
Dean
Oceafography
Redacted for privacy
Deanof Graduate School
Date thesis is presented
/i-ç'-i
/" 7-
Typed by Opal Grossnicjdaus
Grossnicidaus for Ronald
Ronald Keith
geith Reed
for John V. Byrne
ACKNOWLEDGMENT
I am indebted to the late Dr. June C. Pattullo for encouragement and assistance in returning to school after a long absence.
I
would like to thank Dr. William P. Elliott, my major professor,
for initiation into my thesis topic and considerable guidance. His
efforts and skill in helping me to pursue
pursue special
special topics
topics of
of interest
interest
are also 3ppreciated.
appreciated. Dr. Clayton A. Paulson provided helpful
comments on my thesis. Mr. Richard Egami. assisted and advised
me in the analysis of data.
The National Oceanic and Atmospheric Administration made
my return to school possible. I would especially like to thank
Captain W. D. Barbee and Mr. T. V. Ryan of that organization.
My wife, Annemarie, provided constant encouragement, served
as both parents, and solved many problems alone during my absence
at school.
This tnvestigation
investigation was supported by the
the National
National Science
Science Founda.Foundation Grant no. GA-31141.
TABLE OF CONTENTS
I.
1.
II.
III.
INTRODUCTION
1
PREVIOUS PRECIPITATION ASSESSMENTS
3
Extrapolation from Land Stations
The Method of Tucker (1961)
3
RAIN MEASUREMENTS AT TOTEM
Introduction
Types of Gages
Comparison of Gages
Measurements, 1969-1970
Measurements, 1971-1972
Discussion
IV.
ASSESSMENT OF PRECIPITATION AT LIGHTSHIPS
Data and Methods
Results
V.
PRECIPITATION FREQUENCIES
Hourly Frequencies, All Categories
C3tegories
Hourly Frequencies, Rain, Rain Showers, and
Drizzle
Daily Frequencies
Contingency Diagrams
Diurnal Variation of Frequencies
VI.
VII.
VIII.
4
8
8
11
13
20
21
25
28
28
40
40
44
49
50
53
COMPARISON OF
OF RESULTS
RESULTS FROM
FROMLIGHTSJ-IIPS
LIGHTSHIPS WITH
DATA AT OCEAN STATIONS P AND N
55
MECHANISMS
59
OCEANIC AND ATMOSPHERIC IMPLICATIONS
66
BIBLIOGRAPHY
72
LIST OF FIGURES
Pg e
Pge
Figure
3. 1
Location of Totem and the lightships and land stations
used.
9
3. Z2
The spar buoy Totem moored on station.
station.
10
3. 3
A three-inch diameter rain gage without shield.
10
3. 4
A three-inch diameter gage with a shield.
12
3. 5
An eight-inch, tipping-bucket gage suspended in a
4. 1
4. 2
barrel.
12
Location of stations used near Cape Mendocino.
Bathymetry (30-fathom contour) and topography
(1000-foot elevation, dotted lines) is from U.S. D.C.
N.O.S. (1971).
30
Location of stations used near the Columbia River.
Bathymetry (30-fathom contour) and topography (1000S.
O.S.
foot elevation, dotted lines) is from U.S. D.C. N. 0.
(1971).
4. 3
4. 4
4.4
4. 5
31
Location of stations used near the Strait of Juan de
Fuca. Bathymetry (30-fathom contour)
contour) and
and topogtopography (1000-foot elevation, dotted lines, not shown
S. D.
D. C. N. 0. S.
5.
on Vancouver Island)
Island) is
is from
from U.
U. S.
(1971).
32
Mean monthly precipitation (inches) at (a) Blunts
Reef and Eureka, 1954-1966, and at (b)
(b) Columbia
Columbia
River and Astoria,
53-1966. The dashed lines
Astoria, 19
1953-1966.
show the long-term means as of 1966 at Eureka
and Astoria.
35
Swiftsure
Mean monthly precipitation
precipitation (inches)
(inches) at
at (a)
(a)Swi.ftsure
1955-June
say, 1955-June
Bank, Tatoosh Island, and Neah Bay,
1961, and at (b) Umatilla Reef, Tatoosh Island, and
Neah Bay, July 1961-1965. The dashed lines show
the long-term means as of 1965 at Tatoosh Island.
36
Page
Figure
5-1
5. 2
5,
5. 3
5.3
5.4
6. 1
Mean monthly frequency (%) of hours with precipitation
of all categories at (a) Blunts Reef and Arcata and
(b) Columbia River and Astoria, July 1955-June 1958.
Measured precipitation (inches, dashed lines) is shown
for Eureka and Astoria.
42
Mean monthly frequency (%) of hours with precipitation
of all categories at Tatoosh Island and Swiftsure Bank,
July 1955-June 1958. Measured precipitation (inches,
dashed lines) for the same period is also shown.
43
Mean monthly frequency (%) of hours with rain showers
at (a) Blunts Reef and Arcata, (b) Columbia River and
Astoria; and of drizzle at (c) Blunts Reef and Arcata,
and (d) Columbia River and Astoria, July 1955-June
1958.
46
Mean monthly frequency (%) of hours with (a) rain
showers at Tatoosh Island and Swiftsure Bank; and of
(b) drizzle at Tatoosh island
Island and Swiftsure Bank, July
1955-June 1958.
47
Mean monthly frequency (%) of hours with precipitation of all categories at station P (1947-1970), station
N (1946-1968), Swi.ftsure
Swiftsure Bank,
Bank, Columbia River, and
Blunts Reef (July 1955-June 1958).
57
6. 2
Mean monthly precipitation intensity (inches/hour) at
station P, Swiftsure Bank, Columbia River, and
Blunts Reef.
7. 1
Daily precipitation (inches) at (a) Totem
Totem and
and Newport
Newport
coma Beach.
and (b) Totem and
and We
Wecoma
60
LIST OF TABLES
Page
Table
2. 1
3. 1
3. 2
3. 3
3.44
3.
4. 1
5. 1
5. 2
5.3
Assessment of present weather code numbers in terms
of x, y, and z and the
of x, y, and z for sixthe valu.es
values of
hourly reports (from Tucker, 1961).
Comparison of catches of precipitation (inches) at
Hysiop,
Hyslop, 23 October-22 December 1970.
6
15
Comparison of catches of precipitation (inches) at
Southbeach (Newport), 22 December 1970-23 April
1971.
17
Comparison of the maximum catch in a three-inch
Yaquina
Southbeach with those at Newport and Yaquina
gage at
at Southbeach
gage
Bay, 22 December 1970-23 April 1971.
19
Comparison of measured precipitation (inches) at
Totem and Newport (Marine Science Center), 8
October 197
1-14 April 1972.
1971-14
23
Mean annual precipitation (inches) at land stations and
lightships and the percent of nearby land precipitation
occurring at sea.
39
Mean annual frequencies (%) of hours of precipitation
of various types at Blunts Reef, Arcata, Columbia
River, Astoria, Swiftsure
Swiftsure Bank,
Bank, and
and Tatoosh
TatooshIsIsland,
land,
July 1955-June 1958.
48
Contingency diagrams of precipitation events at Blunts
Reef-Arcata, Columbia River - Astoria, and Swiftsure
Bank - Tatoosh Island, July 1955-June
1955-June 1958.
1958. The
number.in
upper left number
in each rectangle is the number of
observations, the upper right number is the percentage
based on total observations, and the underlined values
are percentages based on total precipitation observations.
51
Monthly and annual mean frequencies(%)of precipitation
(all categories) at the lightships according to hour of
occurrence. The hour of observation is listed in
local standard time.
54
Page
Table
7. 1
8. 1
8.2
Summary of the relation of precipitation to the distance
(nautical miles) of a site to an atmospheric low, 11
January-16 March 1970. The first two rows indicate
the percentage of the time precipitation exceeded 0. 30
inches at Totem or the land stations when
when aa low
low was
was
within 0-300 nm, 300-600 nm, 600-900
600-900 nm,
nm, or
or over
over
900 nm (the number of times lows were present in
these distance ranges is shown in the column headshows the
the mean
mean ratio
ratio oL
of rain at
ings). The last row shows
Totem to the land stations when rain exceeded 0. 30
inches on land.
63
Estimates of mean annual evaporation minus precipitation (cm) from Jacobs (1951) and with precipitation
assessed at the lightships (Table 4. 1).
68
from WusUs
Wust's (1936)
S computed
computed from
(%o) S
Values of
of salinity
salinity(%o)
Jacobs
(1951)
and
formula,
and from
fromprepreformula, using
using E-P
E-Pfro?n
fro Jacobs
cipitation at
ci.pitation
at the
the lightships,
lightships, observed
observed salinity at 10 m
(S) from Barkley (1968), and -S°from Jacobs' values
and from the lightship data.
68
RAINFALL OVER COASTAL WATERS
OF THE PACIFIC NORTHWEST
I. INTRODUCTION
Although it is widely accepted that the
the deeper
deeper parts
parts of
of the
the oceans
oceans
are virtually unknown, it is no less true that many processes occurring at the air-sea interface have eluded
eluded solution,
solution. Rainfall over the
back
oceans remains an enigma in spite of serious studies dating back
nearly a century. Our lack of knowledge is caused not so much by
complexities as
as by
by an
ar almost complete lack of data. Pregeophysical complexities
cipitation measurements at sea are still very rare, and many of the
attempts have yielded questionable results. This situation exists
because of the extreme difficulty in obtaining catches from a ship;
the ship motions produce large and rapid changes in the effective
catchment area of gages,
gages, and
and the
the ship's
ship's structure
structuremay
maygreatly
greatlydisdistort the wind field and cause losses in the catch. These problems
W.M.
M. 0.
0. techtechare discussed and recommendations are
are advanced
advanced in
in aa W.
nical note (1962). Although carefully controlled measurements on
ships are desirable, measurements on more stable platforms such
as buoys are to be preferred.
Increased knowledge of oceanic precipitation is vital to further
understanding of a number of oceanic and atmospheric processes.
The oceanographer is concerned with rainfall because of its effect
2
the
on the evaporation-precipitation balance which in turn affects the
surface density and mass distribution whose variations are of interest because of their effect on flow. For example, Jacobs (1951) was
able to account for many features of the
the observed
observed oceanic
oceanic circulation
circulation
from a consideration of evaporation minus precipitation. This vanable is also used to obtain the heat budget
budget for
for aa column
column of
of water
water and
and
for studies of the continuity of salt and momentum. Atmospheric
scientists need increased knowledge of oceanic rainfall because of
the great importance of the heat of condensation of water vapor in
driving atmospheric motions (Jacobs, 1968). This energy first be-
comes available at the site where precipitation is released, so that
the spatial distribution of oceanic precipitation is highly relevant.
Finally, the distribution and seasonal variation of precipitation-
evaporation is significant because it indicates
indicates source
source areas
areas for
for the
the
entrainment of moisture into the atmosphere, which provides an
index of potential precipitation over land.
land.
The present study attempts to better establish
establish the
the tot3l
total precipiprecipitation and its time and space variations over the coastal oceanic area
adjacent to the Pacific Northwest. In addition, differences between
oceanic and land precipitation are stressed. Information is based on
measurements at the Oregon State University buoy (Totem), on esti-
investigamation of rainfall from
from weather
weather reports
reports at
at lightships,
lightships,and
andon
oninvestigation of precipItation
precipitation frequencies at lightships and nearby land stations.
31
II. PREVIOUS PRECIPITATION ASSESSMENTS
Extrapolation from Land Stations
Precipitation values over the oceans have been derived mainly
by direct extrapolation of measurements from gages at coastal
coasta' and
island sites. Generally, the possibility of horizontal gradients from
coast to open sea has not been considered (Jacobs, 1968). One of
the earliest known charts is that of Supan
Supan (1898),
(1898), and
and another
another widely
widely
used one is by Meinardus (1934), cited by Jacobs (1968). On the
other hand, Wust (1936), as cited by Jacobs (1951), derived annual
He related mean
values of
of precipitation
precipitationby
byaadifferent
differentmethod.
method. He
latitudinal values of surface salinity to evaporation minus precipitation from which precipitation amounts were deduced for various
latitudes. The mean latitudinal values, however, do not show areal
differences in detail.
Jacobs (1951)
(1951) was
was aware
aware of
ofthe
theuncertainty
uncertaintyofdirectextrapolatiofl
ofdre ct extrapolation
of coastal precipitation measurements across vast oceanic areas. He
considered that Wust's (1936) latitudinal values of precipitation were
essentially correct, and he used the areal patterns from Meinardus
(1934). In order to obtain agreement between absolute values, a
correction factor of 0. 55 was applied to Meinardus' Pacific values.
His mean annual Pacific chart shows values ranging from about 80
to 150 cm (31 to 59 inches), increasing to the north, in coastal waters
4
prepared seasonal
seasonalprecipitaprecipitaof the Pacific Northwest. Jacobs also prepared
tion charts by consideration of seasonal frequencies of precipitation
observations. Seasonal differences in intensity of prefrom ship'
ship'ss observations.
minor significance.
significance. His results
cipitation were considered to be of minor
indicate that coastal waters adjacent to the western United States
may receive over 40 percent of their annual precipitation in winter
(December, January, and February).
A more recent chart by Drozdov (1953), shown by Malkus
(1962), is based entirely on extrapolation from shore, however,
without considering any increase in precipitation which may be
caused by a "land effect" as suggested, for example, by Haurwitz
and Austin (1944) and Skarr (1955). Drozdov's chart of mean annual
precipitation indicates values of 100 to over 200 cm (39 to more than
79 inches) in the area of this study. A number of researchers tend
to accept Drozdov's values rather than those of Jacobs (1951). As
131), "Since
"Since it is now known that, even
stated by Malkus
Malkus (1962,
(1962,p.p. 131),
in the tropics, all significant rainfall occurs in major synoptic-
scale storms, it is likely that the 'coast effect' on precipitation has
been considerably overrated in the past."
The Method of Tucker (196
(1961)
U
Although Tucker (1961) considered Jacobs' (1951) charts to
be the best available, he noted that considerable uncertainties were
5
caused by (1) deficiencies in Wust's (1936)
(1936) assessment
assessment of
of precipita.precipitation, (2) areal patterns derived from Meinardu& (1934) chart, and
(3) the broad categories of precipitation types used in determining
seasonal charts. Tucker used an entirely different method from
those previously discussed. Amounts were not extrapolated from
land nor were other variables, such as surface salinity, used. A
relation was found between weather report numbers and amounts,
and the factors were applied to reports from ocean station vessels
in the North Atlantic.
Tucker (1961) used the present weather code (ww code, WMO
code 4677), which reserves the numbers 50 to 99 for various types
of precipitation at the time of observation and the numbers 20-29 for
precipitation or fog in the past hour but not at the time of observation. He noted that the crudest method of estimating amounts would
be to relate measured land precipitation to the number of times any
of the code numbers (which are subjectively assigned by observers)
representing precipitation was reported; on the other hand, attempting to solve for 60 different parameters for the 60 code numbers
was not justified because of the very rare occurrence of some types.
Tucker adopted three parameters for assessment of the code numbers
numbers
50 to 99 and considered the contribution of 20-29 negligible. The
parameter x is considered to be representative of light continuous
rain, and y and z are associated with moderate continuous rain and
6
heavy continuous rain, respectively. The assessments, in terms of
these parameters, for the different code numbers are shown in
Table 2. 1.
Table 2. 1. Assessment of present weather code numbers in terms of x, y, and z and the values of
x, y, and zz for six-hourly reports (from Tucker, 1961).
6
7
8
9
x/2
x/2
y/2
x/2
y/2
0
1
2
3
4
55
5S
00
x/2
x/2
x
y
2y
2y
66
x/2
x
y/2
y
z/2
z
xx
7
x/2
x
y/2
y
z/2
z
0
8
x/2
y/2
z/2
x/2
9
2
x
2
x
x
0
z
22
2
0
x/2
x/2
x/2
x,2
2
0
0
z
x = 0. 144 inches
yO.445
inches
y 0.445inches
z = 0. 640 inches
Data from
Data,
fromBritish
British land
land stations
stations were
were used
used with
with equations
equations of the
form ox +3y+'iz5,
+3y+''z5, where
where c,, 3, and
are the number of observed
precipitation and
and66isisthe
themeasured
measuredprepreevents of x, y, and z-type precipitation
cipitation at the land station. An equation was used for each month
at the available stations (a tota,l
total of
of105
105 months
months for
for six
six stations)
stations) and
and
5.
An r. m. s.
x, y, and z were determined by a least square solution. An
error (between estimated and measured precipitation) of about
30%
7
was found for a single monthly mean value. The values of x, y, and
z were then applied to the three-hourly reports from ocean station
live-year period
vessels to estimate precipit4tion
precipitation amounts. For the five-year
used at the ocean stations, errors of 1 3, 8, 3nd 4% were estimated in
the mean monthly, seasonal, and annual values, respectively.
Furthermore, Tucker concluded that the method was not very senseri-
freqiency distributions
sitive to differing frequency
distributions at
at the
the land stations used in
and z.
z.
deriving the values of x, y,
y, and
Tucker
(1961)presented
presentedmean
meanannual
annual and
and seasonal
seasonal precipitation
precipitation
Tucker (1961)
charts for the North Atlantic based on data at ten ocean weather
stations. The patterns deviate considerably from those of Drozdov
(1953)
andamounts
amounts are
are appreciably
appreciably less
less in
in some
some areas.
areas. Because of
(1953) and
large differences between oceanic and nearby coastal and island pre-
cipitation, Tucker concluded that attempts to prepare charts in traused
ditional ways are futile and that oceanic weather data must be used
to establish reliable patterns. Although Tucker's method appears
to be the best available for determining oceanic precipitation
it has
has not been used outside
(Laevastu, Clarke, and Wolff, 1969),
1969), it
the North Atlantic because of the lack of sufficient ocean weather
stations.
III.
RAIN MEASUREMENTS AT TOTEM
Introduction
introduction
As noted previously, actual gage measurements of precipita-
tion at sea are especially needed. This would give an indication of
the validity of the magnitude and distribution presented on the van-
ous precipitation charts. There is still much doubt about the relia-
bility of gage measurements on ships, and a more stable platform
Onebecame
becameavailable
availablein
in 1969
1969 when
when a 185would
woul.d be
be very
very desirable. One
foot spar-buoy (Totem) was moored off the Oregon coast at 45° 04'N,
1240 441
441W
W (see
(see Figures
Figures 3.
3. 1 and 3.2). The buoy was built and de-
ployed by Oregon State University and is moored with a two-point
anchor system; the motion has a long period, and the maximum tilt
does not exceed 100 even in very rough seas (see Neshyba, Young,
and Nath, 1970, for details). Precipitation was measured with gages
at Totem in 1969-1970, and the results have been reported by Elliott,
Egami, and Rossknecht (1971). The buoy was not moored during the
winter of 1970-197
1970-19711 but
but was
was established
established again
again in
in summer
summer of 1971,
and measurements have been made since October of that year. During 1970-1971, when Totem was not in place, comparisons of the
various gages used were made at two land sites.
45°N
40°N
125°W
Figure 3. 1, Location of Totem and the lightships and land stations used.
1 2O°W
10
Figure 3. 2.
The spar-buoy, Totem moored on station.
Figure 3. 3. A three-inch diameter rain gage without shield.
11
Types of Gages
Essentially two different types of gages have been used at
Totem: (1) ordinary collecting gages and (2) a tipping-bucket recording gage. Except for limited use of a ten-inch diameter
diameter gage,
gage, the
the
collecting gages were three-inch diameter gages as shown in Figures
3. 3 and 3. 4. Rainfall entering the receiver drops through plastic
tubing into polypropylene collecting bottles, which are removed to
measure the contents when the buoy is serviced. Two of these gages
have been used on Totem; one was unshielded,
unshielded, and
and the
the other
other was
was
equipped with a wire-mesh, dish-type shield with the top about onehalf inch below
below the
the top
top of
of the
the gage receiver. (These gages were
halfinch
obtained through the courtesy of Mr. Will Shinners, National Oceanic
Atmo spheric Administration,
Administration, Atlantic Oceanographic
and Atmospheric
Oceanographic and
and Meteor
Meteor--
ological Laboratories, Miami, Florida. ) The shield is designed to
reduce the speed of air flow rather than to deflect the wind field,
which is usually attempted.
The tipping-bucket gage also collects rain, but in addition it
records the amount of rain as a function of time. As rain enters
the orifice, it fails into a cup which (when filled) tips downward and
activates an electromagnet. Thus the "tips" registered on an event
recorder indicate the intensity of rain as determined from the rate of
tips of the cups. The problem with using this type of device on a
12
Figure 3.4. A three-inch diameter gage with a shield.
Figure 3. 5. An eight-inch, tipping-bucket gage suspended
in a barrel.
13
platform at sea is that accelerations would cause premature tipping.
bucket gage
gage was
To eliminate this problem, the eight-inch
eight-inch diameter
diameter bucket
23 inches
inches in diamesuspended inside a metal barrel (35 inches
inches high
high by
by 23
than an inch below
ter) by springs so that the top of the gage was less than
5J. A pipe was connected to the
the top of the barrel (see Figure 3. 5).
bottom of the gage with a special weight consisting of a horizontally
mounted disc and four vanes mounted vertically. The barrel is par-
tially filled
ti.ally
filled with
with water
water (to
(to aa level just above the disc and vanes) to
provide damping. Tests indicated that the disc-vane arrangement
was quite effective in preventing premature tipping from vertical
attached rigidly to
and horizontal motions of the barrel, which was attached
the buoy.
calibraThe tipping-bucket mechanism was adjusted and a calibra-
tion factor was derived for converting number of tips to inches of
precipitation.
Comparison of Gages
In October 1970 comparisons of various gages were made at
Hyslop, a farm with flat terrain seven miles
miles northeast
northeast of Corvallis,
maintains aa station
Oregon. The National Weather Service also maintains
State University
there with a recording gage. A total of six Oregon State
gages were used: (1) two eight-inch tipping-bucket gages suspended
in barrels, one with damping water and one without water; (2) two
three-inch unshielded gages; and (3) two three-inch shielded gages.
14
These gages were positioned in a north-south line about Z5
25 feet long;
about 1155 feet
the Weather Service gages were about
feet east
east of
of the
the center of the
and there were no other obstructions to wind or rain. The barline, arid
rels were placed on platforms about four feet high, and the three-inch
gages were mounted on metal strips, which were attached to the tops
of the platforms and projected outward. Wind speed and direction were
sensed at the same site and recorded on a strip-chart recorder.
Results from the tipping-bucket gages were compared to those
from the recording Weather Service gage for the period 23 October9 December 1970. For 40 comparative measurements of daily precipi-
tation, the standard deviation of tipping-bucket values from the
Weather Service results
results was
was ±0.
±0.027
0Z7 inches;
inches; thus
thus the random error
for these gages at 95% confidence limits is ±0.05 inches. Precipitation collected by the various gages is indicated in Table 3. 1. Means,
standard deviations, two standard deviations (the presumed random
measurement error at 95% confidence limits), and the percentage
error (ratio of measurement error to mean) are shown for each
period. The mean percentage error is 6%, and the standard devia-
percentage error
error is
is ±±2%.
tion of percentage
2%. There
There is not a consistent trend of
greater catches with the shielded gages; in fact, in six cases average
values are greater for the unshielded gages. Also, in six of eight
cases results from both gages mounted in the barrels were less
than the mean, but the differences were generally very small.
15
Table 3. 1. Comparison of catches of precipitation (inches) at Hyslop, 23 October-22 December
1970.
Type of
gage
23
23 Oct.
Oct. 26 Oct. 9 Nov.
9 Dec.
Dec.
Nov. 9
12 Nov. 16
16 Nov.
Nov. 24
24 Nov. 30 Nov.
22 Dec.
start
0.98
1.33
1.33
0.49
j.84
1.84
1.95
2.83
2.97
start
Fight-inch
Eight-inch
(in barrel
without water)
0.97
i.33
1.33
1.35
0.49
2.00
1.97
2,85
2.99
start
1.01
1.40
1.36
0.52
2.04
2.08
3.01
3.24
3,24
start
1.03
1.42
1.39
0.51
1.88
2.04
2.91
--
Three-inch
(shielded)
start
0.89
1.35
1.38
0.52
187
1.87
2.02
2.80
.08
3.08
Three-inch
(siielded)
(shielded)
sta't
start
0.97
1.35
1.40
0.52
1.85
2.02
2.87
3.04
0.97
1.36
1.37
0.51
1.92
2.Oj
2.01
2.88
3.07
Eight-inch
(in barrel
with water)
Three-inch
(unshielded)
Three-inch
(unshielded)
mean
s
0.O48
±0.048
2s
2s
±0.09
0.09
%error
9
0.037
0.
037
±0.07
55
±0.026
0. 026±0.014
±0.014
±0.05
0.0S
4
0.03
±0.03
6
±0. 108
±0.042 ±0.074
0. 074 ±0.108
0.085
0.
085 ±0.042
0.17
±0.17
9
0.08
±0.08
4
±0.15
5
±022
±0.22
7
16
It should be noted, however, that these measurements were made
under light wind conditions; the highest daily average was only 12
miles/hour.
mites/hour.
On 22 December 1970, three gages (one eight-inch tipping-
bucket type mounted in a barrel with water, one three-inch unshielded,
and one three-inch shielded) were installed at Southbeach (Newport,
Oregon) and were maintained until 23 April 1971. The three-inch
gages were attached to poles, which were crossed at right angles
and mounted to the corner posts and braces of a metal fence at about
six-feet height. The barrel containing the tipping-bucket gage was
placed atop a six-foot concrete structure about 15 feet north of the
three-inch gages; the top of the barrel was above the fence. In
addition, winds were recorded at this same site. Data at the Southbeach site are compared with those from Weather Service stations
(U. S.D.
S.D.C.N.
(U.
C. N. 0.
0. A.A.
A. A. E.D.S.,
E. D. S.,1970-1971)
1970-197 1)atatthe
theYaquina
Yaquina Bay
Bay Coast
Guard Station (north side of Yaquina Bay) and at Newport (10th and
Eads Streets, elevation about 150 feet); both sites are within two
miles of the beach location.
Table 3. 2 is a comparison of the precipitation measured (and
recorded) at Southbeach. The "weighted wind" at Southbeach is
also shown for those periods when hourly wind speeds and recorded
precipitation were both available. The weighted wind was computed
by multiplying the wind speed (at an hour when precipitation occurred)
Table 3. 2. Comparison of catches of precipitation (inches) at Southbeach (Newport), 22 December
1970-23 April 1971.
Type of
gage
12 Mar. 23Apr.
l2Mar.
22 Dec.
6 Jan.
22 Jan.
26 Jan.
11 Feb.
18 Feb.
5 Mar.
start
3.07
7. 52
2. 12
1.72
1. 17
1.79
1.20
1.75
1.88
--
--
Eight - inch
Eight-inch
tipping -bucket
Measured
Recorded
--
2.20
--
4.22
10.08
2.90
2.62
1.58
3.06
2.79
12.13
12. 13
4.53
9.60
2.83
2.65
1.52
3.25
2. 59
2.59
11.10
32
24
27
35
26
46
36
22
19.
19.66
17. 2
3. 31
Three-inch
9.45
(unshielded)
Three-inch
(shielded)
tt
H
Deficit(%)
Deficit
(%)
(8U gage catch
compared to
compared
to max.
max.)
)
Weighted Wind
(knots)
22.0
ratio of
by the rati.p
of precipitation
precipitation for
for that hour to the daily total. These
values were then summed to obtain a daily total weighted wind,
which was then adjusted in a like manner and summed to derive
the weighted wind for the desired period. It is immediately appar-
ent from Table 3. 2 that the gage in the barrel caught less rain than
Onthe
theother
otherhand,
hand,there
thereisisno
noconsistent
consistent differdifferthe smaller
smaller gages.
gages. On
ence between the three-inch gages; variations from the mean range
from 1 to 4%, which is somewhat lower than the measurement errors
inferred Lrom
from the
the data
data at Hyslop. The deficits from the eight-inch
gage ranged from 24 to 46%. This effect is assumed to result from
thar Hyslop; the deficits,
the much higher wind speed at Southbeach than
however, do not appear to be directly related to wind speed. Perhaps
losses only occur above a critical wind speed. The large size of the
barrçl could be expected to provide a disturbance to wind flow which
would be much greater than for the three-inch gages.
A comparison was made of the maximum catch at Southbeach
with precipitation measured at the Weather Service stations at
Yaquina Bay Coast Guard Station and in the town of Newport. Some
uncertainty exists, of course, because these sites may receive
different amounts of rain than the beach site; considering the elevation (about 150 feet) of the Newport site,
site, however,
however, and
and the
the more
more
inland location of the Coast Guard station, one would expect greater
greater
precipitation than at Southbeach.
3.
The results are
are shown
shown in
in Table
Table3.3.3.
Table 3. 3. Comparison of the maximum catch in a three-inch gage at Southbeach with those at
Newport and Yaquina Bay, 22 December 1970-23 April 1971.
5 Mar.
23Apr.
12 Mar. Z3Apr.
22 Jan.
26 Jan.
11 Feb.
18 Feb.
4.53
10.08
2.90
2.65
1.58
3,25
2.79 12.13
(2) Newport
5.24
5. 24
11.85
3.50
3.05
1. 50
1.50
3.63
3. 18
12.04
(3) Yaquina Bay
5.52
5. 52
12. 19
12.19
3.63
--
1.90
3.60
---
--
Location
(1) Southbeach
22 Dec.
start
Deficit(%)
Deficit
(%)
[(2) compared
to (1)]
[(3) compared
to (1)]
to(1)]
6 Jan.
14
15
17
11
-5
10
12
-1
18
17
20
--
17
10
--
--
'.0
20
Generally, more precipitation was measured at the Weather Service
stations, with Yaquina Bay receiving slightly more than Newport. The
differences, however, are not great; the maximum deficit is only 20%
and the mean is 12%.
The highest loss did occur during the period of
the highest computed weighted wind (22-26 January), but there are
too few values (and not enough knowledge about wind at all sites) to
determine if this was a casual factor. Although exact calibration
factors cannot be reliably derived, the three-inch gages at Southbeach
caught amounts of rain that were not greatly
greatly different
different from
from the
the inshore
inshore
stations; the differences probably result from possibly higher winds
at Southbeach or from increased rainfall at the more inland sites.
Measurements, 1969-1970
In November 1969 a tipping-bucket gage mounted in a barrel was
installed on the west side of Totem at a height of about 35 feet, depending on the ballast in the buoy and the height of the superstructure
above sea level. During two of the measurement periods in early
1970, a ten-inch non-recording gage was also used; it was mounted
on a pole and suspended at the same height
height as
as the
the other gage. The
1969-1970 measurements were reported by Elliott, Egarni,
Egami, and
Rossknecht (1971). The catches at Totem were compared with five
Oregon coastal sites from Astoria to Newport. The results were
quite interesting; all coastal sites reported at least twice as much
21
rain as at Totem for all four periods when the rain catch was collected
at the buoy.
The most typical.
typical figure
figure was
was about four times as much
presented
rain on the coast. Elliott, Egami, and Rossknecht also presented
supporting evidence from weather reports taken in the area by Oregon
State University' s R. V. Yaguina. Further, they concluded that winds
were not consistently higher at Totem than at coastal sites. (An
apparent wind effect on the tipping-bucket gage was subsequently
noted from the comparison at Southbeach, however.
however, Increasing the
catch at Totem by about 1. 5 though would not seriously alter the
conclusions drawn; precipitation for the first two periods would still
be about three times greater on the coast, and the comparison for
the last two periods, which showed the smaller differences, was
based on catches from the unobstructed ten-inch gage) Thus the
conclusion remains that appreciably less rain fell at sea than on the
coast that winter. The recordings from the tipping-bucket gage
in chapter
provide additional information; they will
will be
be discussed
discussedin
VII, however, rather than here.
Measurements 1971-1972
After Totem was remoored in 1971, rain gages were again
instafled. They consisted of a tipping-bucket gage as before, which
installed.
was placed on the southwest corner of the buoy deck, and a shielded
and an unshielded three-inch gage, which were mounted on metal
22
strips that were attached to and projected from the southeast corner
of Totem. The receivers of the three-inch gages were about ten feet
above the top of the
the barreL
barrel.
The same
that the tipping
same type
type instruments
instrumentsasasatatTotem
Totem(except
(exceptthatthe
tipping--
bucket gage was not placed in a barrel) were installed at the Marine
Science Center of Oregon State University in Newport in a semi-
meeting Weather
Weather Service
Service specifispecifienclosed location near the ground meeting
cations (Richard Egami, personal communication). Attempts were
made to record tips at Totem during this period, but they were not
not
successful because of repeated failures
failures of
of the
the paper
paper drive
drive mechanism.
mechanism.
1-1972 measurements
measurements to date are presented
Results of the 197
1971-1972
large gage
gage in
in the
the
in Table 3. 4. Once again, it is apparent that the large
barrel did not catch as much rain as the small gages. The deficits,
especially the one for the period 6 January-7 March, were somewhat
greater than most of those determined at Southbeach; a complicating
factor is that the barrel was not mounted outboard as in 1969 but was
placed on the buoy deck, and gas tanks lashed nearby may have
further obstructed the air flow. It is interesting that in all cases
the tipping-bucket gage at Newport caught slightly more rain (the
differences are less than 10%) than the three-inch gages, and the
not consi,stently
consistently produce higher readings. Conshielded gage did not
verseLy,atatTotem
versely,
Totemthe
the shielded
shielded three-inch
three-inch gage
gage did catch more rain
every period. The deficits for the unshielded gage range from 3 to 9%
Table 3.4. Comparison of measured precipitation (inches) at Totem and Newport (Marine Science
Center), 8 October 1971-14 April 1972.
Gage
Totem (tipping-bucket)
Totem (3t1 unshielded)
Totem (3t1 shielded)
8 Oct.
18 Nov.
17 Dec.
7 Mar.
14 Apr.
5. 87
3. 79
3. 34
6 Jan.
- -
start
start
5.26
5.
26
5. 36
3. 53
8.64
8.
64
5.22
5.
22
5.58
5.
58
5.91
3.84
9.03
9.
03
5. 37
40
58
38
8
4
3
*
Deficit (%)
gage catch compared
to max.
max.)
(811
(8t1
Deficit (%)
(3tt unshielded
(31t
unshielded compared
to 3" shielded)
Newport (tipping-bucket)
Newport
(3t1unshielded)
unshielded)
Newport(3t'
Newport (3" shielded)
% Precipitation at Sea
(ratio of max. catch at
6
9
start
11. 31
11.31
16.60
16.60
"
10.25
15.13
15.
13
10.72
10.03
"
10. 37
14.86
9.93
49
36
36
36
26. 31
11. 58
22.71
2Z.71
23.45
10.46
10.66
10. 66
34
46
Totem to that at Newport)
Rainfall caught by the tipping-bucket gage was not collected on 17 Dec.
The value listed for this gage under 6 Jan. is for the period 18 Nov. -6Jan.
(J
24
with a mean of 6%; these are the first data we obtained that clearly
show this effect.
As was indicated by the measurements two years before, the
catch at Totem was appreciably less than that at Newport. (All
gages show this effect, but the sea-land ratios shown in Table 3. 4
are based on maximum values at Totem and Newport.) During midwinter only about one-third as much rain was obtained at Totem as
on land; during fall and spring the ratios increased to almost one-half.
This tendency toward greater disparity
disparity in
in precipitation
precipitation amounts
amounts at
at
sea and ashore during winter is also evident in the analysis of lightship data to be presented. It should be stressed that an anomalous
amount of rain did not fall at the Marine Science Center compared
with other coastal sites. For the period from 8 October 1971 to
6 January
(U.(U.S.D.C.
S.D. C. N.N.O.A.A.
0. A. A. E.
D. S.,
January 1972,
1972,when
whendata
dataare
areavailable
available
E.D.S.,
197 1-1972), amounts listed for coastal sites used by Elliott, Egami,
and Rossknecht
Rossknecht (1971)
(1971) are:
are: Astoria
Astoria--29.7;
29.7;Nehalem
Nehalem- -61..
61.8;
8; Otis 50. 4; and
and Newport
Newport - 36.
36. 99 inches. Rainfall for this period was 39. 6
inches at the Marine Science Center but was only 15.3 inches at
Totem.
Thus Totem appears to receive only one-fourth
one-fourth to
to one-half
one-half
as much rain as coastal sites from the Columbia River to Newport.
25
Dis cus sion
In addition to instrument comparisons, other tests were made
to confirm the validity of the catches at Totem. First, a polypropylene bottle (like the ones used to hold the catches) was filled with
water and left on the beach for a month,
month, but
but no
no difference
difference in
in quantity
quantity
was found.
Thus, evaporation of the catch at Totem would not appear
to be a factor. Second, all of the samples have been analyzed for
salinity in case extremely high waves might have forced sea water
or spray into the receivers. No values great enough to significantly
affect the volumes were detected, however. The gage comparisons
at Totem and at Southbeach strongly indicate that the tipping-bucket
gage did not catch the maximum possible amount of precipitation
under strong wind conditions. The effect was not noted at Hyslop
(winds less than ten knots) but became apparent at Southbeach for
winds in excess of 17 knots; there is no
no evidence,
evidence, however,
however, that
that the
the
effect increases with an increase in wind speed. The mechanism
producing precipitation losses from the gage mounted in the barrel
is probably turbulence caused by the large size of the barrel as an
then displace
displace rain
rain
obstruction to the wind; the resulting eddies
eddies i-nay
may then
drops so that they do not all fall into the collecting cup, which is only
a small proportion of the total area of the top of the barrel. The
conclusions drawn regarding the low precipitation at sea compared
26
to the coast are not altered by problems with this system, however.
This instrument was not directly compared with coastal catches in
as noted
noted above,
above, any
any correction
correction factors
factors suggested
suggested by
1971-1972, and,
and as
later gage comparisons could not account for the large sea-land
i
1970 for
for the
the two
two periods
periods when
when data from
differences found in 19691969-1970
this gage were used.
The comparisons of the three-inch gages at Southbeach did not
yield unequivocal proof of the catching efficiency of these gages, but
the results are strongly suggestive. The small deficits observed
for these gages, compared to those farther inland, are not proof
that the strong winds on the beach reduced the catch in these gages.
They may just as well indicate that the actual precipitation at Newport
and Yaquina Bay was indeed higher because of the more inland locations and the higher elevation at Newport. The fact that the shielded
gage at Totem consistently caught more rain than the unshielded one,
which was not the case in other places, is of interest. The meaning
of this is not entirely clear, but it may be that losses in these small
gages do not occur until very high wind velocities are encountered.
Thus, perhaps slight losses (a mean of 6%) occurred in the unshielded
gage under very high winds at Totem, but the mesh screen reduced
seems implausimplausthe effective wind velocity as it is designed to do. It seems
ible that
ibl.e
thatserious
serious losses
losses could
could have
have occurred
occurred in
in both
both three-inch
three-inch gages
because the differences between the gages are fairly consistent, but
27
the winds were probably very much higher at some times than others.
Further, W. M. 0. (1962) concluded that a small gage suspended at
a height of 10 m well away from obstructions would catch essentially
as much rain as on the ground even in very high winds. Thus, the
evidence suggests that essentially such conditions held at Totem
with the three-inch gages and that the measured amounts are reliable
estimates of the rain at sea.
!13
IV. ASSESSMENT OF PRECIPITATION AT LIGHTSHIPS
Although the measurements at Totem have provided what is
believed to be a reliable index of oceanic precipitation at that site, it
would be desirable to compare these results
results with
with independent
independent assessassessments. Also, the results at Totem provide no information on areal
differences in precipitation over coastal waters. It appeared that
data from lightship stations off the coast might help satisfy both these
needs, and it was decided to try to assess precipitation amounts from
weather reports using the method derived by Tucker (1961).
Data and Methods
mid-1950's,
Since the mid1950's, weather
weather reports
reports have
have been
been made
made at
at sixhourly intervals in WMO Code 4677 (present weather, ww) at the light-
ship stations maintained by the U. S. Coast Guard as aids to navigation.
The data are on file at the National Climatic Center (U.S.
Commerce, National
National Oceanic
Oceanic and
and Atmospheric
Atmospheric AdminAdminDepartment oL
of Commerce,
istration), Asheville, North Carolina and are available in punched
card form (International Marine Surface Synoptic Observations, Card
Deck 128) for varying periods for the Pacific Coast lightships. The
stations analyzed were Blunts Reef, Columbia River, Swiftsure Bank,
and Umatilla,
Umatilla Reef
Reef (which
(which replaced
replaced Swiftsure
Swiftsure Bank
Bank in
in Juty
July 1961); their
locations,
loca,ti.ons,plus
plusthose
thoseof
ofthe
theland
landstations
stations used,
used, are
are shown
shown in Figures
4. 1-4.3. Print-outs of the data available in punched card form
were obtained from the National Climatic Center for the following
1966), Columbia
1966),
periods; Blunts
Blunts Reef
Reef (1954(1954-1966),
Columbia River
River (1953(1953-1966),
Swiftsure Bank (1955-June 1961), and Umatilla Reef (July 1961-1965).
1961-1965).
The original plan was to establish a relation between measured
precipitation and present weather reports at the west coast land
stations and use these values to assess the rainfall at the lightships.
That is, the values would have been derived
derived as
as Tucker
Tucker (1961)
(1961) did
did for
for
British land stations. Unfortunately, it was not practical to do this.
Although synoptic observations of present weather (WMO Code 4677)
are made (usually four times a day) at the land stations, these data
are not routinely punched (William Bartlett, personnel communica-
tion) as the hourly observations are. The hourly observations (WBAN
Hourly Surface Observations, Card Deck 144), however, do not contam
tam
precipitation listed in the ww code, and the categories used are
too broad to allow correlation with ww code numbers. The synoptic
observations could have been used, but the cost of obtaining paper
copies of the data for a reasonable period was prohibitive.
Consequently, the values of the parameters (x, y, and z)
derived by Tucker (1961) have been applied directly to the precipi-
tation frequencies at the lightships. Although this approach is less
than ideal, Tucker's analysis indicates that it should have sufficient
validity. First, he found that the errors in precipitation estimates
30
41°N
1
Sta.
Sta.
:',
\ ::,
'\
(e12O7
//(el207
_
;;
31fm
//
/
I
'
I
I
S.
I
I
Blunts
__ --;
i.
i.
\_,____
\_,____
\
Reef
lightship
Cape
Midocino
Mizidocino
\
-- --
'
/
I
..
\
/__
\
/1,__
,___'_
('_
-
.-s
I
\
-''-
I
S
.1
(5
-
\
-
-
I)
\ N\S
S
_'S
I
';
-
/ ,
I
-
40°N
II
II
I
I
I
'\
I'
SS
I
124°W
Batiiymetry (30-fathom
(30-fathom
stations used
used near
nearCape
CapeMendocino.
Mendocino. Bathymetry
Figure 4. 1. Location of stations
contour) and topography (1000-foot elevation, dotted lines) is from U. S.D. C.
N.O.S. (1971)
31
31
\o1bia
ver4
yer
Columbia
ver
lightship
\\
Astoria
t k
:
Sta. (el.8')
\
3Ofm
46°N
46N
's)
I'
s
-'"5-I
rs
r'
"S
4'
I
I
's,
*
F
_,
I
/
I
-'
-
;
-
I
/
il\
I
1
24°W
124°W
Figure 4. 2. Location of stations used near the Columbia River. Bathymetry
(30-fathO&ncontour)
(30-fathIrn
contour)and
andtopography
topography(1000-foot
(1000-footelevation,
elevation, dotted
dotted
lines) is from
from U.S.
U.S.D.C.
D.C. N.O.
N.O.S.
5. (1971).
32
32
49°N
'S
/
Vancouver
I.
Vancouver 1.
Swiftsure
Bank
-.
lightship
Strait of
-
/
\
Ju:de
Fuca
3uandeuca
Tatooshl,6
TatooshI,4.T8ay
Sta.(ei.c!'
(el.\Sta.
'
-.
Sta.'e1. 15').
Sta.eL
1OV)
101')
3Ofm
r
.'; -
Umatilla
Umtilla\
,*
S-.,,*
f\.
Reef
'\
lightship
(
/
-s
- -_
-
,-
48°N
S.-
-_:,
;
I
125°W
125°w
1 24°W
Bathyrnetry
near:the Strait
Strait of
of Juan
Juan de
de Fuca.
Fuca. Bathymetry
Figure 4.3. Location of stations used near:the
(30-fathom contour) and topography (1000-foot elevation, dotted lines,
not shown on Vancouver Island) is from U.S.D.C. N.O.
N.O.S.
5. (1971).
33
(and
at a station were quite insensitive to the
the groups
groups of
of stations
stations (and
their frequency distribution of weather reports) used in deriving the
values of x,
y,
even when
when stations
stationswith
withaasignifisignifiand z. This was so even
cant orographic effect on the east coast of Britain were included, and
the r. m. s. difference between measured and estimated precipitation
at an east coast Canadian station was only slightly greater than for
the British stations used. Second, Tucker concluded that the standards of British and American observers on ocean weather ships were
very similar and that significant errors did not arise from possible
differences in reporting. Finally, the general weather off northeast-
em Europe is quite similar to that off the Pacific Northwest; thus
precipitation in both regions is associated with cyclonic systems.
Table 2. 1 shows the assessments of present weather in terms
of x, y, and z after Tucker (1961); Tucker's values for x, y, and z
are 0. 144, 0. 445, and 0. 640 inches respectively for six-hourly
observations. Occurrences of precipitation at the lightships were
tabulated for each month of record, and the values were converted
to precipitation amounts using the assessments and the values for
x, y,
x,
y, and
and z. Monthly and annual means were then determined.
It
shou'd be noted that the monthly means have not been adjusted for a
This presumpresumseasonal variation in
in error
error found
found by
by Tucker
Tucker(1.961).
(1961). This
ably results from secondary maxima of precipitation in spring and
in late summer. Such features are not present in the stations used
34
here, however, and the maximim corrections were only approximately 20% for months when there is very little precipitation off
the Pacific Northwest.
Results
Mean monthly precipitation at the lightship stations is shown
in Figures
Figures 4.
4. 44 and
and 4.
4.5.
5.1 Measured precipitation at nearby land sta-
tions(U.S.D.C.
tions
(U.S.D.C. WB.,
W,B.,1953-1966
1953-1966and
and TJ.S.D.C.
TJ.S.D.C. W.B.,
W.B., 1955-1965
for Neah Bay) is included for comparison. The land stations are,
of course, somewhat different in relative locations to the lightships,
and their elevations vary. Blunts Reef is four nautical miles from
the nearest iand
land and is 27 and 38 nautical
nautica' miles from Eureka and
Arcata, respectively. Columbia River is seven nautical miles from
land and 15 nautical miles from Astoria.
Swiftsure Bank is six
nautical miles from land (Vancouver Island)
island) and is 17 and 21 nautical
miles respectively from Tatoosh Island
Is]and and Neah Bay; comparable
figures for Umatilla
tJmatilia Reef are 5, 14, and 15 nautical
nautical miJes.
miles. The
elevations at Eureka and Arcata are 79 and 207 feet respectively, but
air flow from the ocean is not blocked by intervening hills. Astoria
(elevation eight feet) is almost due east of the Columbia River light-
ship, and the terrain between it and the coast is very flat. Tatoosh
Island is a flat topped rock, 100 feet high and almost 0. 2 miles in
diameter, located
located 0.
0.44 miles northwest of Cape
Cape Flattery
Flattery (U.
(U. S.
S.D.
D. C.
Month
Month
M
15
M
Jj
I
I
S
I
N
J
M
M
J
S
N
15
15
I
Eureka
Blunts Reef
110
10
-8
.-
7
0.
U
0
0
4.
4.
0.
0
precipitatIon (inches) at (a) Blunts Reef and Eureka, 1954-1966, and at (b) Columbia River and
Figure 4.4. Mean monthly precipitation
1953-1966. The dashed lines show the long-term means as of 1966 at Eureka and Astoria.
Astoria, 19S3-1966.
(J
u-I
U-'
Month
M
M
S
JJ
N
N
J
J
20
C)
NealiBay
Tatoosh Island
IIII
MM
Month
J
N
S
J
o Neah Bay
Tatoosh Island
Swiftsure Bank
L
Umatilla Reef
15
15
2
U,
7/
4,
4)
4)
..c
C)
U
/\
,/
U
C)
I
0
10
.2
1C
4..
4-.
4-.
0
U
C)
U
4)
S
A
0
I
_____
(b)
(a)
I
(1
Figure 4.5
4.5. Mean
Meanmonthly
monthlyprecipitation
precipitation(inches)
(inches)atat(a)
(a) Swiftsure
Swiftsure Bank,
Bank, Tatoosh Island, and Neah
Neah Bay,
Bay, 1955-June
l955-June 1961, and at
at
(b) Umatilla Reef, Tatoosh
Tatoosh Island,
Island, and
and Neah
Neah Bay,
Bay, July
july 1961-1965.
1961-1965. The
The dashed
dashed lines
lines show
showthe
thelong-term
long-termmean
meanas
as
of 1965 at Tatoosh Island.
(J.)
(J-)
C'
37
37
C & G. S., 1951). Consequently, a more inland station, Neah Bay,
has also been included; although its elevation is only 1 5 feet, winds
from the west or south might cross elevations of 500 feet or more.
All of the comparisons (using Tucker's method) between light-
ships and land stat.ons reveal striking differences in precipitation
(Figures 4.4 and 4. 5). Although the seasonal trends are quite simiin summer),
[ar at all stations (showing maxima in
in winter
winter and
and minima
minimain
rainfall at the lightships is much less than over land. The long-term
thevanvanmeans at the land stations are very similar to the means for the
ous periods of data used. This suggests that the periods were ade-
quate to produce representative monthly means at the lightships pro-
viding that the errors inherent in the assessments are not too great.
Tucker (1961) estimates that the r.m. s. percentage error for a fiveyear period is 13% in monthly means and 4% in annual rainfall; except
for Umatilla Reef, the periods used here are all greater than five
years.
(Originally, the analysis for Columbia River was performed
for the period 1953-1959 and results were compared with those at
Astoria. After this was done, however, Columbia River data for
1960-1966 became available and were obtained, and a separate analy-
sis was performed for this period. The results used here are for
the combined period (1953-1966), but findings from the separate
analyses are of interest. Precipitation at Astoria was 8% less during
1960-1966than
thanfrom
from l953to
1953to 1959,
1959,
1960-1966
and the
the same percentage reducand
tion for the later period was found at Columbia River. The striking
similarity of land-sea ratios for these two periods suggests that the
relatively short periods at Swiftsure Bank and Umatilla Reef can
also be used with
with some
some confidence
confidence for
for assessments
assessments by
by Tuckerts
Tuckers
method.)
Figures 4.
It is apparent from Figur'es
4. 44 arid
and 4. 5 that the precipitation
differences between lightships and land stations are greater during
(Differencescomparable
comparable to
to those
those in
in absolute
winter months.
months. (Differences
the winter
values are not consistently found in the ratios, however, partly
because of the relatively great percentage effect of the measurement
error on the slight precipitation during summer. ) Another interesting
feature is the large differences in rainfall at Tatoosh Island and Neah
Bay.
The amounts at Tatoosh Island are intermediate between those
at Neah Bay and the lightships; thus this small offshore island appears
to have significantly greater rainfall than over coastal waters, but
precipitation is less than that at more inland Neah Bay.
Table 4. 1 presents the mean annual precipitation at land stations and lightships for the various data periods. Rainfall at sea
is generally less than half that over land if one excludes the results
for Tatoosh Island. The results are very similar tothose found from
gage measurements at Totem, even though the lightships are all less
than eight miles from the coast and Totem is 30 miles offshore.
39
Although annual rainfall increases appreciably (at land and sea)
from Cape Mendocino
Mendocjno to the Strait of Juan de Fuca, it is uncertain
if the ratio of sea to land precipitation does because of the differing
results for Swiftsure Bank and Umatilla Reef. At any rate, the
differences are not large,
large, and
and itit appears
appears that
thatvery
verysimi
simUar
Jar ratios
ratios
may exist over a vast coastal area.
Table 4. 1. Mean annual precipitation (inches) at land stations and lightships and the percent of
nearby land precipitation occurring at sea.
Precipitation
Precipitation % at sea
Period
Land station
1954-1966
1954-1966
Eureka
Eureka
38
Blunts Reef
15
40
1953-1966
Astoria
72
Columbia River
25
35
l9SS-June
l955-June 1961
Neali Bay
Neah
97
Swiftsure Bank
49
SO
50
Tatoosh
Tatoosh Island
Island
79
Swiftsure Bank
49
62
Neab
Neah Bay
98
Umatilla Reef
37
37
38
Tatoosh
Tatoosh Island
Island
75
Umatilla Reef
37
37
50
50
July 1961-1965
Lightship
40
V. PRECIPITATION FREQUENCIES
A readily available piece of information from the marine observations is the frequency of precipitation. In order to compare these
results with those on land, printouts of hourly surface observations
were obtained from the National Climatic
Climatic Center for Arcata (rather
Astoria,
than Eureka for which the data were not
not available
available on
on tape),
tape), Astoria,
and Tatoosh
land. Although the record length chosen is not great,
Tatoosh Is
Island.
with cost considerations it was felt to be of sufficient duration to
to
show major features.
Further, the period (July 1955-June 1958) was
chosen to agree closely with the long -term means at the sites. Differences in mean monthly precipitation for this period and the long term
means were all less than two inches, and most were less than one
inch.
May
The main differences were that the months
months of
of April
April and
and May
were drier than normal at all sites, June and July were wetter than
Novemusual at Astoria and
and Tatoosh
Tatoosh Is
Island,
land, October was wetter and Novern-
ber drier than normal at all stations, and greater than normal rain
occurred at Arcata during December-February.
Hourly Freguencies,
Frequencies, All Categories
The first frequency analysis performed was for all types of
precipitation (including freezing or frozen forms) grouped as a
single class. This was done by determining the percent of
41
precipitation occurrences for each month from the hourly observa-
tions at the land stations and the six-hourly reports at the lightships.
Monthly and annual means for the period used were then derived.
Figures 5. 1 and 5. 2 show the mean monthly frequency of hours with
precipitation for Blunts Reef-Arcata, Columbia River-Astoria, and
Swiftsure Bank-Tatoosh Island. The features are strikingly similar
to those present in precipitation amounts (Figures 4.4 and 4. 5) that
were based on longer series of data. The seasonal trends for these
four stations are similar in all cases, except what appears to be an
unusuafly low frequency at Columbia River in January. Frequency
unusually
half that
that on
on land,
land, and
and the
the differof rain at sea is generally less than half
data. There is good
ences increase in winter as noted for other data.
agreement between features in the plots of frequency and of measured
precipitation for this period. Especially apparent are the dry May
and wet June and the wet October and dry November in both plots at
Astoria. Thus the frequency of total precipitation appears to be a
good general indicator of the amount of rainfall, which implies that
differences in intensity of rainfall are not highly significant to mean
monthly (or longer) data.
Annual mean frequencies of hours of precipitation (all types)
were also derived for these stations for the three-year period. The
resulting values (in percent) are: Arcata - 14, Astoria - 22, Tatoosh
8, and Swiftsure
Island - 23;
Reef - 4, Columbia
Columbia River
River - 8,
23; Blunts Reef
Island
M
M
Month
M
50
so
I
I
M
Month
S
N
I
I
o Astoria
Columbia River
o Arcata
Blunts Reef
40
40
30
30
15
is
0
t)
ti)
I
:
20
-
/
20
I
U
10
//
10
10
-0
ET1Ia)
3
0 -'----------
5
S
0
Figure 5.
5.1.
(b) Columbia
Columbia
Meanmonthly
monthlyfrequency
frequency(%)
(%)ofofhours
hourswith
withprecipitation
precipitationof
ofall
allcategories
categories at
at (a)
(a) Blunts
Blunts Reef and Arcata and (b)
1. Mean
Measured
precipitation
(inches,
dashed
lines)
is
shown
for
Eureka
and
Astoria.
(inches, dashed lines) is shown for Eureka and Astoria.
River
July 19SS-June
1955 -June 1958.
River and
and Astoria,
Astoria, July
19S8.
N?
43
Month
J
M
M
J
S
J
5U
su
3 Tatoosh Island
Swiftsure Bank
40
15
V
U
0
c)
0
U
20
10
S
n
0
Figuxe 5.
Figw?e
5. 2.
2. Mean
Mean monthlyfrequency
monthly frequency(%)
(%)of
ofhours
hours with
with precipitation
precipitation of
of all
all
categories at Tatoosh Island and Swiftsure Bank, July 1955-June
1958. Measured precipitation (inches, dashed line) for the same
period is also shown.
44
Bank - 12. The ratios of frequencies (sea to land) are: Blunts Reef -
0.29;
Arcata, 0.
29; Columbia
Columbia River - Astoria, 0.36; and Swiftsure Bank Tatoosh Island, 0. 52. Comparable ratios of mean annual precipitation
amounts (Table 4.1) are as follows: Blunts Reef - Eureka, 0.40;
Columbia River - Astoria, 0.
0.35;
35; and
and Swiftsure
Swiftsure Bank
Bank -- Tatoosh
Tatoosh Island,
0. 62.
The general agreement of ratios for frequencies and amounts
seems too good to be fortuitous, and it lends further credence to the
assessments of amounts from present weather reports at the lightships.
Hourly Frequencies, Rain, Rain Showers, and Drizzle
Additional examinations of frequency were made by grouping
types of precipitation into three rough categories: rain, rain showers-,
and drizzle. This was done in the following way.
For hourly obser-
vations at the land stations, the codes R, R, and
R+
as rain, RW, RW, and
RW+
were classed
were grouped as rain showers, and
L , L,
L,
L, and
and L+ were listed as drizzle (see U. S. D.C. N.
N. C.
C. C.,
C., 1970).
1970).
At the lightships, present weather codes 58-59, 60-65, and 91-94
rain; 80-82
80-82 were
were listed
listed as
as rain
rainshowers,
showers and 50-55
were classed as rain;
and 58-59 were considered drizzle. (The codes 58-59 are described
as drizzle and rain. Since they have been used in two categories, as
done by
byN.W.S.C.
N. W. S.C.E.D.,
E. D.,1971,
1971,the
thesum
sumofoffrequencies
frequenciesof
ofthese
these types
types
may not exactly equal the frequencies of all categories.
45
Mean monthly frequencies of hourly precipitation as rain
The plots
5. 4.
showers and drizzle
in Figures 5. 3 and 5.4.
drizzle is
is shown
shownin
of rain are not shown because of their extreme similarity to Figures
5. 1 and 5. 2.
5. 1
There is an increase in rain showers from Arcata to
Astoria; of interest, however, is that frequencies at Tatoosh Island
are reduced appreciably from those at Astoria. The most striking
feature is the virtual absence of rain shower reports at sea. It is
not known if this reflects a real difference between the lightships and
land stations or if rain showers at sea are listed as some other form
of precipitation. The rain shower trends on land, however, show the
same seasonal variations as seen in the precipitation amounts and
the frequencies of all categories of precipitation. The distributions
of frequency of drizzle
drizzle are
are considerably
considerably different
differentfrom
fromother
otherp1ots
plots.
First, there is little seasonal trend apparent, except possibly for
mid-summer maxima at Arcata and Tatoosh Island. Second, there is
less disparity between frequencies at sea and on the coast than seen
in all the other comparisons. The above tends
tends to
to support
supportTucker1s
Tucker's
(1961) conclusion that a greater percentage
percentage frequency
frequency oL
of total
total precipiprecipi-
tation at sea falls as
as drizzle
drizzle than
than on
on land;
land it does not support his suggestion, however, that the same may be true for showers.
Table
Table 5.
5. 11 summarizes the information on mean annual frequen-
cies for
des
for the
the various
various classes
classes of precipitation. The first column does
not equal the sum of the last three in all cases because of rounding,
Month
Month
20
JJ
-*---#---+-4-----
M
M
M
;M
Jj
S
N
J
20
o Arcata
Blunts Reef
(a)
(a)
0
UC)
10
10
w
0)
0)
0
0
Month
Month
J
r
1
J
3
S
N
J
i
o Astoria
Mtoria
10
>.10
-
I
Blunts Reef
o
110
(d)
(c)
10
0
Columbia River
01
Figure 5.3. Mean monthly frequency (%) of hours with rain showers at (a) Blunts Reef and Arcata, (b) Columbia River and
Astoria; arid
and of
of hours
hours with
with diizz1e
diizzle at
at (c)
(c) Blunts
Blunts Reef
Reef and
and Arcata,
Arcata, and
and (d)
(d) Columbia
Columbia River
River and
and Astoria, July 1955June1958.
0'
C.'
47
Month
10
0
10
I
Figure-S.4. Mean monthly frequency () of hours with (a) rain showers at
Figure-5.4.
Tatoosh Island and Swiftsure Bank; and of (b) drizzle at Tatoosh
Island and Swiftsure Bank,
Bank, July
July 1955-June
1955-June 1958.
1958.
of hours
hours of
of precipitation
precipitation of
of various types at Blunts Reef, Arcata,
(%)of
Table
1. Mean annual frequencies (%)
Table 5.
5.1.
Columbia
River, Astoria, Swiftsure Bank, and
July1955-June
1955-June 1958.
and Tatoosh
Tatoosh Island, July
1958.
Columbia River,
All categories
Blunts Reef
Rain
Rain showers
Drizzle
4
Z2
0
2
14
7
4
3
8
5
0
2
Astoria
22
11
8
2
Swiftsure Bank
12
6
0
4
Tatoosh Island
23
113
3
5
55
Arcata
Columbia River
49
inclusion of codes 58-59 in two categories, and neglect of freezing
or frozen precipitation in the last three columns. Rain is seen to
account for about half of all precipitation frequencies, and the sea-
land ratios are much as for total frequencies and for amounts. As
noted before, rain showers were virtually never reported on the
Astoria.
ocean, and there is a noticeable peak in shower activity at Astoria.
Drizzle events are similar at land and sea and appear to be most
frequent off the Strait of Juan de Fuca.
Daily Frequencies
In an attempt to gain possibie
possible information
information aboiit
about the mechanisms
mechanisms
systemsiproducirig
(or systemsIp
roducing precipitation at coastal and lightship sites,
another factor was examined.
examtned. The number of days with precipitation
lightships and
was determined from six-hourly observations at the lightships
land stations, and the percentage frequencies were computed. This
is a common practice used to derive information from land stations
frequencies of
of days
days with
withprecipiprecipi(Jacobs, 1968). The mean annual frequencies
tation were: Arcata - Blunts Reef (33%-1l%);
(33%-I1%); Astoria - Columbia
River (46%-20%);
(46%-ZO%);and
andTatoosh
TatooshIsland
Island -- Swiftsure
Swiftsure Batik
Bank (47%-28%).
(47%-Z8%).
The ratios (mean annual) for these stations pairs are respectively:
0. 33, 0. 44, and 0. 60, which are only slightly greater than mean
annual. ratios
annual
ratios based
based on
on hours
hours with precipitation (all classes). It is
believed, however, that these values do not reflect reality. A
50
"rainy
observations
rainy event" for each day is determined from four observations
both at land and sea; the hourly frequencies are much lower at sea,
however, and the chance of observing a rainy day is therefore less
at sea. That is, even if as many rainy days occur at sea as on land,
the chances of recording them is less than half as good. Also, the
limited recordings at Totem and Wecoma Beach (to be discussed in
Chapter VII) suggest that there is not much disparity between days
with precipitation at sea and land.
Contingency Diagrams
Contingency diagrams were prepared for the station pairs
Blunts Reef-Arcata, Columbia River-Astoria, and Swlftsure
Swiftsure BankTatoosh Island in an effort to find relations between precipitation
events at sea and on land. This was done by noting and tabulating
precipitation at either station of the pair and listing the corresponding type of precipitation (or lack of it) at the other station. The
observations used at both stations were those at the time of the sixhourly observations at the lightships.
The present weather codes
used for the categories of precipitation were the same as previously
noted, except that codes 58-59 were classed only as drizzle in order
to avoid listing two events at one station with only one at the other.
2.
The results are presented in Table 5. Z.
Various features (such as frequency of drizzle at a lightship
Table 5. 2. Contingency diagrams of precipitation events at Blunts Reef-Arcata, Columbia River-Astoria, and Swiftsure Bank-Tatoosh Island,
July 1955-June 1958. The upper left number in each rectangle isis the number of observations, the upper right number is the percentage based on total observations, and the underlined values are percentages based on total precipitation observations.
Arcata
Arc at a
Nn rIiri
i\Tri
DIin
3341
84.6
99
2.5
a
30
0.8
49
4.9
26
4
0.1
1
LO
1.0
0.9
34
4
.
.
-o
0.3
13
2.11
2.
83
9.1
1.0
40
76.8
3039
4)
0.2
6
3.1
0
S
0.2
0.8
0.9
Cl,flwer
(lSwPrC,
t'211S
253
6.4
32
0.8
0.1
140
6.55
6.
22
0.6
2.4
3.5
3
256
28.2
27.8
7
3.7
34
rain
rlrl77lp
,Iri77lp
25.7
0.5
19
Nn vain
Mn
rain
4.0
157
33:0
0.7
0.7
chnwrc
hnwerc
5.1
202
16:2
4) -
Astoria
rit,
clri77lp
dviz,1p
3.5
34
0.9
0-
ci,
5
I 21
6.6
0.2
4.3
0.1
0
0
0.1
5
0.1
2
C,
S.
a
21
21
2.
1
2.1
0.3
3.7
0.1
5
0
C
C
(6 t
(6
C
-c
S.
Tatoosh Island
dri,1e
dr1e
rain
6.3
135
6.3
257
26.0
13.7
1.3
5.5
37
S.
23
0.6
0.9
.
11
8
0.3
172
2.1
1inwer
163
163
4.0
22
06
2.2
4.2
22
0.6
2.2
17.4
1.1
Total precip. obs. - 909
16.5
8.7
2.3
37
3.7
C
-C
C,
0
2.
2.
0.1
2A
Cd,
a
2
Total observations - 3948
Total precip. obs. - 611
No riin
Mn
rii,
C
3077 75.6
54
0
2.
Total observations - 3952
a
3.7
15.4
0
0
P.
2.
0
5S
21
Total observations - 4064
Total precip. obs. - 987
0.1
0
2.
0
4
21
0.1
52
Considstation) that have already been discussed are evident here. Consid-
erable additional information is available, however. The most freqtient
quent (based on total precipitation observations) groupings in order
at Arcata-Biunts
Arcata-Blunts Reef are rain, rain showers, and drizzle on land
and
with no precipitation at sea followed by rain at both places and
drizzle and rain at sea only. At Astoria-Columbia River the order
is essentially the same, except that rain showers on land only is the
most frequent group, and rain at both locations assumes more importance. At Swiftsure Bank-Tatoosh Island, the order is somewhat
drizzle on
on 1and
land
different: most frequent is rain, rain showers, and drizzle
only followed by rain at both places, drizzle at sea-rain on land, and
drizzle and rain at sea only. At Arcata-Blunts Reef there is more
than at
at the
the other
other pairs; precipiindependence in precipitation events than
tation occurs simultaneously only 14.
14. 8%
8% of
of the
the time
time that precipitation
and Swiftsure Bankis occurring, whereas at Columbia River-Astoria
River-Astoria and
Tatoosh Island values are 26. 7 and 34. 4%. At Astoria-Columbia
River and Tatoosh Island-Swiftsure Bank simultaneous rain occurs
atone. At
more frequently than the totals of precipitation at sea alone.
Astoria, rain showers assume considerable
considerable importance,
importance, and at the
relatively frequent,
northermost station pair drizzle at sea becomes relatively
with much of it accompanied by rain on land.
53
Diurnal Variation of Frequencies
A final frequency analysis was made to determine if a diurnal
variation was present in the occurrence of precipitation; the complete
data periods obtained
3). At
obtained for
for the
the lightships
lightships were
wereused
used(Table
(Table5.5.3).
Blunts Reef there is convincing evidence for a diurnal variation in
hours than
than
the annual means; appreciably less rain falls at 10 and 16 hours
during hours of darkness. This trend is also evident in the monthly
means, except for May-June and August-October. At Columbia
River there is evidence
afternoon, but some
evidence or
or the
the least
least rain
rain in
in midmid-afternoon,
monthly means do not show this feature. There is only a very slight
suggestion of a diurnal variation at Swiftsure Bank, and the monthly
means show little consistent trend. Nearby Umatilla Reef, however,
appears to have a definite variation with less rain falling in the afterof June. Thus, in gennoon; the only exception to this is the
the month
month of
during
eral there appears to be a tendency for more precipitation during
hours of darkness or early morning. A similar analysis was per1968. The
formed for selected land stations by M. C. P. N. R. B. C., 1968.
results for the coastal locations of Brookings
Brookings and
and Nehalem,
Nehalem, Oregon
Oregon
and Clearwater, Washington also show somewhat of an increase in
frequency during darkness, principally from 00 to 06 hours.
54
Table 5.3. Monthly and annual mean frequencies (%) of precipitation (all categories) at the
lightships according to hour of occurrence. The hour of observation is listed in
local standard time.
Hour
Hour
Jan,
Jan, Feb.
Mar.
Apr.
Mar. Apr.
May June
July
Aug.
Sept,.
Sept.. Oct.
Nov.
Nov. Dec.
Blunts Reef (1954-1966)
04
30
37
37
33
33
30
62
33
43
50
16
33
31
10
10
24
22
20
15
25
15
22
29
10
10
16
21
21
16
17
17
15
15
23
23
22
25
8
11
0
20
25
22
23
22
28
25
24
30
19
15
33
29
20
44
24
25
25
Annual mean: O4hr-33%;
O4hr-33%; lOhr-21%;
lOhr-21%; 16hr.209o;
16hr20%; 22hr-26%
22hr-26%
Columbia River (1953- 1966)
03
26
30
33
28
39
28
30
30
35
26
24
25
33
09
09
24
21
21
22
22
30
20
26
24
24
7
25
25
24
24
15
19
19
18
25
16
27
20
13
27
20
23
20
21
31
30
27
17
25
19
26
28
40
31
28
28
23
23
Annual mean: 03 hr-29%; 09 hr-24%; 15 hr-20%; 21
21hr-27%
hr27%
1961)
Swiftsure
Swiftsure Bank
Bank (1955-June 1961)
03
25
24
24
29
23
37
24
45
32
32
31
27
27
28
28
26
09
28
23
27
29
29
23
23
27
16
22
24
30
28
23
15
22
27
24
26
25
28
16
22
18
24
20
22
21
25
27
20
22
16
20
20
24
23
27
18
24
28
Annual mean: 03 hr-27%;
hr-27%; 09
09 hr-26%;
hr-26%;15
15hr-23%;
hr-23%;21
21hr-23%
hr23%
Umatilla Reef (June 1961-1965)
03
27
34
36
31
31
34
37
29
29
29
29
28
24
09
09
28
26
23
27
34
19
20
20
21
39
24
19
24
15
15
20
20
17
14
14
17
3
28
17
6
10
24
17
17
20
21
25
24
27
25
31
19
25
44
44
23
23
24
36
32
32
21 hr -- 28%
28%
hr-25%; 15
15 hr-17%;
hr-17%; 21
Annual mean: 03 hr-30%; 09 hr-25%;
55
VI. COMPARISON OF RESULTS FROM LIGHTSHIPS
WITH DATA AT OCEAN STATIONS P AND N
It seems of interest to compare, in at least a cursory manner,
information from the lightships with results from a summary of data
data
ocean stations
stationsPPand
andNN(N.
(N.W.S.C.
E.D.,
from ocean
W. S. C. E.
D., 1971).
1971). Ocean station
300 N,
N,
P is located at 50°
500 N,
N, 145°
145° W
W and
and N
N isis at
at 30°
1400W; they are thus
and
approximately 850 nautical miles northwest
northwest of
of Swiftsure
Swiftsure Bank
Bank and
Figure
1000 nautical miles southwest of Blunts
Blunts Reef,
Reef, respectively.
respectively. Figure
6. 1 shows the mean monthly frequency of hours with precipitation at
ts apparent that ocean station P
these stations and the lightships. It is
receives much more frequent precipitation
precipitation than
than the
the other
other stations.
onlyabout
about
summer precipitation
precipitation is
is only
Also of interest is the fact that summer
four percent less than in winter. At station N, precipitation is less
annual mean
mean
than at Swiftsure Bank but more than at Blunts Reef; its annual
is approximately the same as Columbia River. It too has a smaller
annual variation than the lightships.
The relatively frequent occurrence of precipitation is difficult
to reconcile with data reported by Allen
Allen (1963).
(1963). He reported the
results of a measurement program at station P where rain gages
for nine years. The annual
were mounted on ocean weather vessels
vessels for
inches with annual values
mean of measured precipitation was Z66 inches
ranging from 21 to 29 inches; further, Allen, after considering the
be too
factors involved in the catches, concluded that they might be
56
almost the
the same
same as
as at
at
large. This annual mean precipitation is almost
Columbia River and only slightly greater than half that at Swiftsure
Bank. Since frequencies and amounts were both available for the
lightships and station P, the precipitation intensity was computed and
mean monthly values are shown in Figure
Figure 6.
6. Z.
2. The data periods for
determining frequency are the same as for Figure 6. 1; the amounts
are based on longer periods at the lightships. Some of the variation
in intensity at the lightships likely results from the short data period
used for frequency as well as the very low values for summer amounts
apparent,howhowand frequencies (which are used as divisors). It is apparent,
ever, that the intensity at station P is much less than at the lightships.
Annual mean values are 0. 013, 0. 05Z,
052, 0. 039, and 0.039 inches/hour
for station P, Swiftsure Bank, Columbia River, and Blunts Reef,
respectively. Thus if the catch data are reliable, the intensity of
rain must be on the average only one -fourth of that at the lightships.
Such a great disparity perhaps casts some doubt on the rain measurements reported by Allen (1963). This appears to be a problem worth
iurther investigation, and it suggests the possible uncertainties in
further
attempting to extrapolate distributions across sizeable distances
over the oceans.
A final comparison was made with the diurnal variation at the
before, most
mostoL
of the
lightships and at station P and N. As noted before,
lightships showed some evidence for increased rainfall at night or
57
Month
Month
J
40
M
M
I
M
M
J
S
N
J
I
o Station
Station PP
o Station N
Swiftsiu,e
Swiftsuxe Bank
L
L.
30
£
Columbia River
Blunts Reef
20
10
o!____ ___I
0
i
Figure 6. 1. Mean monthly frequency (%) of hours with precipitation of all
categories at station P (1947-1970), station N (1946-1968),
Swiftsure Bank, Lojumbia
uoiumbia River, and Blunts Reef (July
(July 19551955June 1958).
Month
0.
o. 2
a
J
M
M
J
S
N
j
Station P
t
Swiftsure Bank
A
Blunts Reef
Columbia River
0.1....,
Q)
0
I
0L
0'-
Figure 6. 2. Mean monthly precipitation intensity (inches/hour) at station P, Swiftsure
Bank, Columbia River, and Blunts Reef.
early morning. Although not completely consistent, the same trend
is apparent at both ocean stations, but it is much better developed in
summer than in winter months.
VII. MECHANISMS
The results of three virtually independent analyses have been
presented and discussed. They were based on gage catches of rain
rain
at Totem during two 'Trainy seasons," conversion of present weather
codes at lightships to estimates of precipitation amount, and examinatton of the frequencies of precipitation at lightships
tion
lightshtps and land stations.
The results are in rather remarkable agreement; they indicate that
less than half as much rain falls over coastal waters as on land, with
the added indication that this occurs mainly because it rains more
hours over land. Little information has been presented about why
this happens, however. Although attempts were made regularly to
obtain records
records of
of tips
tips from
from the
thetipping-bucket
tipping-bucketgage,
gage,data
datawere
wereobohtamed only during two short periods in 1970. It is believed, however,
that these data, in conjunction with other information, provide hints
as to possible mechanisms for the disparity between oceanic and
coastal rain.
During the period 11 January-4 February 1970 continuous
records were obtained at Totem, and from 6 February to 16 March
they were obtained at Totem and (with an identical gage and recording
Comparisons of daily
system) at
system)
at Wecoma
WecomaBeach
Beach(see
(seeFigure
Figure3.3.1).
1). Comparisons
precipitation at Totem and land stations are shown in Figure 7. 1.
(The values indicated for Newport are daily measured precipitation
Day
Day
Jan. 1970
10
Feb. 1970
Feb.
20
1
3
33
o
Newport
Totem
US
US
C)
U
o
0
L
(a)
\
oi
0 H
s;
0
Figure 7. 1. Daily precipitation (inches) at (a) Totem and Newport and (b) Totem and Wecoma Beach.
Mar.
61
from U.S. D.C. N. 0. A. A. E. D. S., 1970-197 1, with readings at
Totem adjusted to the same day basis, rather than recordings. )
No
attempt was made to correct the readings at Totem or at Wecoma
Beach for the possible losses with the bucket-gage during this period;
thus the absolute values shown in Figure 7. 1 may be in error, but
relative differences with time would be valid. The total precipitaNewport -- 17.
17.22
tion converted from these measurements is: (1) Newport
inches; Totem - 4. 2 inches; and (2) Wecoma Beach
9.9 inches,
Totem
Totem - 2. 6 inches. The differences are too large to be accounted
for by any correction factor so far indicated, and the ratios for the
two station pairs are essentially equal.
A very interesting point emerges from the comparison in Figure
7. 1.
Over both periods it rained about one-fourth as much at sea as
on land.
This relation is far from uniform over time, however. On
some days it rained about one-half as much or more at sea as ashore
(12-15 January, 24 January, 15-16 February for example), but on
other days less than two-tenths as much rain fell at sea (e. g., 16-23
January, 26 January, 30 January, 6 February, and 6 March). In an
attempt to understand what might cause such large variations, daily
weather maps (U. S.D.
S.D. C.
C. E.
E. S.
S.S.
S. A. E. D. 5.,
S., 1970) were examined.
It appears that, at least in a qualitative manner, the different ratios
of rain (sea to land) were associated with different atmospheric sys-
terns; thus, when relatively large amounts of rain fell at sea, an
tems;
atmospheric low was usually in the vicinity, but heavy precipitation
62
on land and light rain at sea was frequently associated with frontal
passages. The distance from Totem to an atmospheric low was
measured for each day, and this was compared with the precipitation at sea and at Newport or Wecoma Beach. Table 7. 1 gives the
percent of the days Totem and the land stations received more than
0. 30 inches of rain in relation to distance from the low. Even with
the inexactness and uncertainty of determining lows, a trend seems
greater than
than 0.
0. 30
30 inches
inches at
at
clear. Most of the precipitation events greater
Totem occurred when a low center was within 300 nautical miles of
the site; on the other hand, there seems to be little relation of rainfall on land to the location of a low. The last row in Table 7. 1
shows the mean ratio of rain at Totem to land when precipitation
on land was in excess of 0. 30 inches; thus these values are derived
from the number of days producing the percentages shown in row
two. A trend is again apparent. When Totem receives heavy precipitation relative to land, low centers are usually nearby.
Since hourly values were available at sea and on land during
the period of record at Wecoma Beach (6 February -16 March), the
frequency of hours with precipitation was computed at both sites.
The values are 10% for Totem and Z0%
20% for Wecoma Beach, much
Freas the results of the other frequency comparisons discussed. Fre-
quency of days with precipitation was also determined; values were
61 and 77% respectively for Totem and Wecoma Beach. (It should
Table 7.1. Summary of the relation of precipitation to the distance (nautical miles) of a site to an
atmospheric low, 11 January-16 March
March 1970.
1970. The first two rows indicate the percentage
low
of the time precipitation exceeded 0. 30
30 inches
inches at
at Totem
Totem or
or the
the land
land stations
stations when
when aalow
900 nm
nm (the number of times lows
was within 0-300 nm, 300-600 nm, 600-900 nm, or over 900
column
headings). The last row
the
were present in these distance ranges is shown in
shows the mean ratio of rain at Totem to the land stations when rain exceeded 0. 30 inches
on land.
0-300 nm
(no. = 8)
300-600nm
(no.
16)
600-900
(no.z18)
(no.18)
>900
(no.23)
Totem
50
12
0
A
Newport or
Wecoma Beach
50
38
44
35
Mean ratio
(sea to land)
0. 67
0. 34
0. 1 5
0. 20
0'
0tJ)
(J
64
also be noted that there was no difference
difference in
in days
days with
with precipitation
when Totem was compared to Newport). The fact that there is no
appreciable difference in days with precipitation on land and at sea
lends support to the conclusions in Chapter V that these daily freweather
quencies are not reliable when computed from six-hourly weather
reports.
The above indications allow some tentative conclusions to be
reached. First, when it rains on land during a given day, precipi-
tation usually also occurs at Totem. It rains only about half as
many hours at sea on this day, however. Second, it rains relatively
vicinity; although it
more at sea when an atmospheric low is
is in
in the
the vicinity;
also rains at sea when frontal systems (not lows) produce land pre-
cipitation, the amounts are usually relatively quite small. These
facts suggest that the mechanisms producing
producing greater
greater land
land precipita-.
precipita-
tion are less important in the presence of a low.
low, Thus a low may
result in greater convective activity at all locations, whereas rain
occurring with a frontal passage may be light at sea but is greatly
enhanced by frictional processes over land. Such processes do not
appear to be dependent on the orographic effects of high mountains
but may occur over short distances across low-lying land. As noted
before, some such magnification process appreciably increases
precipitation on a feature as small as Tatoosh Island. These conclusions are in opposition to the statement by Malkus (1962, p. 131)
65
significant rainrainall significant
"Since
Since it is now
now known
known that,
that, even
evenin
inthe
thetropics,
tropics, all
fall occurs in major synoptic-scale storms, it is likely that the
'coast effect' on precipitation has been considerably overrated in
appears,rather,
rather, that
that aa very
very sharp
sharp gradient
gradient of precipithe past.
past." ItItappears,
tation amounts may be present right at the shoreline.
VIII. OCEANIC AND ATMOSPHERIC IMPLICATIONS
Do the conclusions drawn have relevance to oceanic areas other
than the coastal Pacific Northwest? Tucker's (1961) analysis of rainfall over the North Atlantic indicates appreciably
appreciably less
less rain
rain compared
compared
to land as was found here. Large areas of the global ocean are sub-
ject to precipitation associated with cyclonic atmospheric systems.
Thus, one might expect similarities in rainfall
rainfall over
over the
the ocean
ocean adjacent
adjacent
to the Pacific coasts of northern North America and southern South
America and the Atlantic coast of northern Europe. Extension of
precipitation assessments to sizeable
szeab1e oceanic regions (other than the
North Atlantic) is presently limited by the lack of ocean weather sta-
tions and reliable precipitation measurements
measurements from
from ships.
ships. A
A possible
possible
interim solution might be the assessment
assessment of
of precipitation
precipitation amounts
amounts
from reports by merchant and naval vessels in otherwise data sparse
regions.
The results of this study indicate that mean annual precipitation
over coastal waters of the Pacific Northwest is appreciably less than
previous assessments (Jacobs, 1951; Drozdov, 1953). The amounts
derived are considered to apply from very near the beach to at least
30 miles offshore, but the outer limit is probably much greater than
this. Mean annual values range from 15 inches (39 cm) at Cape
Mendocino to 43 inches (109 cm) at the Strait of Juan de Fuca (mean
67
of two stations). Essentially 40% of the annual precipitation at all
stations fails during the winter months of December, January, and
February in close agreement with the ratio derived by Jacobs (1951).
The mean annual amounts for this area from Jacobs, however, are
about 80 to 1 50 cm. Thus they are appreciably greater than indicated
by this study, but they have the same range (70 cm) over the area.
The values of Drozdov (1953) are even greater than those of Jacobs
as noted in Chapter II.
It is apparent then that previous estimates of evaporation minus
it
minus
precipitation (E-.P)
(E-P) must
must also
also be seriously in error unless the assessments of evaporation are also erroneous in a compensating fashion.
Such is not indicated in a review by Malkus (1962). She concluded
that the values derived by Jacobs (1951) with the transfer formula
method are fairly reliable and are in close agreement with more
recent computations (for example, Budyko, 1956). Since Jacobs
(1951) maps are quite direct to work from, E-P values have been
estimated for the vicinity of Cape Mendocino, Columbia River, and
Strait of Juan;de Fuca, and his values are compared with those com-
puted with the
th present
presentprecipitation
precipitation estimates
estimates (Table 8. 1). It should
be stressed that E-P based on precipitation at the lightships is
appreciably closer to Jacobs' values than they would be to values
derived from the precipitation estimates
estimates of
of Drozdov
Drozdov (1953)
(1953) or
or to
to
earlier maps such as those of Meinardus (1934). Even so, the
Fstimates of mean aimual
annual evaporation minus precipitation
precipitation (cm)
(cm) from
from Jacobs
Jacobs (1951)
(1951) and
and
Table 8.1, Estimates
with precipitation assessed at the lightships (Table 4. 1).
From
Jacobs
(1951)
precipitation
at lightships
Cape Mendocino
Mendocino
Cape
-20
+20
Columbia River
-85
-10
Strait of Juan de Fuca
-90
-50
Location
(1936) formula,
formula, using E-P from
Wust's (1936)
Table 8.2. Values of salinity O) S computed from Wu.st's
Jacobs (1951) and from precipitation at the lightships, observed salinity at 10 in
m
S-S from Jacobs' values and from the lightship data.
(S)
(S) from
from Barkley
Barkley (1968),
(1968), and S_S
(A)O)
Location
S0
(Jacobs,
1951)
5
S
0
(from
lightships)
S
(from
Barkley,
1968)
S-S
S-S
S-S
S-S
(from
Jacobs,
1951)
(from
lightships)
0
0
Cape Mendocino
33.36
34.04
32.8
-0.6
-1.2
Columbia River
32.26
33.53
32.5
+0.2
-1.0
Strait of Juan de Fuca
32. 17
32.85
32.0
+0.2
-0.8
-.0.8
Two principal
principal concl.uconcludisagreement of
of values
values in
in Table
Table8.8.11isislarge.
large. Two
disagreement
diluted by rain
sions follow:
follow: (1) oceanic waters in this region are
are not
not diluted
sions
as much as previously
previous'y assessed; and (2) there is less heat gain in the
atmosphere from precipitation than was thought.
Jacobs (1951) compared the observed surface salinity to salini-
Wust'ss determinations
determinations
ties derived from Wust'
(1936) formulas;
formulas; since
since Wust'
Wust'ss (1936)
were based on zonal averages and were little affected by circulation
features, the difference in these two quantities is a reflection of the
valueswere
wererereJacobs' values
effects of advection and dilution from land. Jacobs'
computed (using his E-P values in Table
Table 8.1)
8.1) from Wust's formula
difficult to
to scale
scale the
the values
values precisely
precisely from Jacobs'
because it was difficult
S is the computed
, where S
Sk,
chart. Wust's formula is S0=K(E-P) + S
k where o
salinity, K is a constant (0. 0170), and Sk is 33. 70%o for this area.
The values of S using Jacobs' E-P results and those based on pre-
cipitation determined at the lightships in this study are shown in
Table 8. 2.
The observed salinity, S, was determined from Barkley
(1968); winter values were used to eliminate the effect of summer
upwelling in part of this region. The
- S differences in Table 8. 1
with the
the values
values
agreement with
based on Jacobs' E-P values are in general agreement
from his original chart, even though he used summer salinities
derived from few data.
The salinity differences,
- S, from Jacobs' results and
results of this study are strikingly dissimilar, except that both show
70
more negative values toward the south. Values based on the older
precipitation estimates would imply no dilution effects in the northern
reveal features that are in
area. It is believed that the newer values reveal
better accord with present-clay
present-day oceanographic knowledge. The increased negative values to the south show clearly the effects of the
California Current flowing along the edge of the higher-salinity central
water, which strongly
strongiy influenced Wust' s zonal averages. On the other
of Juan
Juande
de Fuca, and the
hand, this flow is less marked off the Strait
Strait of
differences are less negative. Jacobs' values, however, are positive
which
indicates that
that there is no diluting mechanism in the region as
which indicates
far south as the Columbia River. The negative values from this
study though suggest appreciable dilution, which agrees with our
knowledge of coastal runoff in the area.
The significance of reduced (compared
(compared to previous estimates)
values of precipitation over the oceans is perhaps even more important to atmospheric dynamics than it is to oceanic processes. The
greatest portion of energy available for maintaining the general circulation
lation.of
ofthe
theatmosphere
atmosphere isis derived
derived from
from the latent heat gained by
the atmosphere during the fall of precipitation
precipitation to the earth' s surface
(Jacobs, 1968).
Therefore, the sites (and amounts) of oceanic rain-
fall
fail are highly important to understanding and forecasting atmospheric
events. Jacobs (1951) presents maps of pc' the heat actually gained
by the atmosphere, which is the sum of the sensible heat conduction
71
(or convection) and the heat gained during precipitation. Mean annual
shown respecrespecvalues of about 150, 200, and 225 g cal/cm2/day are shown
and the Strait
tively for near Cape Mendocino, the Columbia River, and
Juan de
de Fuca;
Fuca; approximately
approximatelythese
thesesame
samevalues,
values, however, appear
of Juan
to apply over a much larger oceanic area.
area.
Basing the Q pc values on the present precipitation estimates
would yield rates of only about one-half Jacobs', except off the Strait
of Juan de Fuca where the revised value would be approximately 175 g
cal/cm2/day. In addition to the equational regions and the western
sides of oceans, the extreme northeast Pacific was listed by Jacobs
as an area with large gains of heat to the atmosphere; the present
results suggest that this is less so than previously thought. (It
should be noted that it is possible that Q pc is not in error as much
as would be implied from the precipitation differences; this would
require deficiencies in previous assessments of the conduction term,
however.) Jacobs also noted that the global
global oceans
oceans are
are not
not sources
sources
of moisture for rain
rain over
over land
land in
in summer.:
summer. Since the E-P amounts
thought, however, revisions
appear to be greater in places than was
was thought,
of some of these conclusions may be warranted
warranted in
in light
light of the important implications.
72
BIBLIOGRAPHY
Allen, W. T. R. 1963. Precipitation measurements at ocean weather
station UPH. 21 p. (Canada. Department of Transport. MeteorCIR-3870, TEC-476)
TEC-476)
ological Branch.
Branch. CIR-3870,
ological
Ocean.
Oceanographic atlas of the Pacific Ocean.
1 56fig.
fig.
University of
of Hawaii
Hawaii Press.
Press. 20 p., 156
Honolulu, University
Barkley, R. A.
1968.
National
of Commerce,
Commerce, National
Bartlett, William. 1972. U. S. Department of
National
Climatic
Oceanic and Atmo
spheric Administration,
Administration,
Atmospheric
Center, Asheville, North Carolina. Personal communication
(telephone). Jan. 21, 1972.
earth's surface. GidBudyko, M.
M. 1.
I. 1956. The heat balance of the earthts
rometeorologicheskoe izdatel' stvo. Leningrad. 255 p. (Translated by N. A. Stepanova. Translation distributed by U. S.
Weather Bureau. Washington, D. C., 1958. 254 p.)
skoi
Norskoi
Drozdov, 0. A. 1953. Annual amounts of precipitation. Nor
Atlas, Vol. II. Chart 48b. (Cited in: Malkus, J. S. 1962.
Large-scale interactions. In: The sea, ed. by M. N. Hill.
131)'
Vol. I. New York, John Wiley. p. 131)
Egarni, Richard. 1972. Research Assistant, Oregon State University, School of Oceanography. Personal communication.
Corvallis, Oregon. June 23, 1972.
Elliott, W. P., Richard Egami and Gary Rossknecht. 1971. Rainfall
at sea. Nature 229:108-109.
Haurwitz
Naurwitz B., and J. M. Austin. 1944. Climatology. New York,
McGraw-Hill. 410 pp.
Jacobs, W. C. 1951. The energy exchange between the sea and
atmosphere and some of its consequences. Bulletin of the
California,
alifornia,
Scripps Institution of Oceanography, University of C
6:27-1 22.
6:27-122.
The seasonal apportionment of precipitation
over the ocean. In: Eclectic climatology, ed. by Arnold Court.
63-78.
Corvallis, Oregon State University Press.
Press. p.
p. 63-78.
1968.
73
1969. Oceanic part
of the hydrological cycle. 71 p. (World Meteorological
Organization, Report no. 11)
Laevastu, T., L. Clarke and P. M. Wolff.
Wolff;
The sea,
sea, ed.
In: The
Malkus, J. S.
5. 1962. Large-scale interactions. In:
by M. N. Hill. Vol. I. New York, John Wiley. p. 88-294.
Meinardus, W. 1934. Eine neue Niederschlagskarte der Erde.
(cited in:
in: Jacobs,
Petermanns
Peterrnanns Geographische Mitteilungen 80:1-4 (cited
The
seasonal
apportionment
of
precipitation
over
W. C. 1968.
the ocean. In: Eclectic climatology, ed. by Arnold Court.
65)
Corvallis, Oregon State University Press. p. 65)
Meteorology
Committee. Pacific Northwest Basins Commission.
Commis sion.
Meteorology Committee.
states,
Columbia
basin
Climatologicalha.ndbook.
handbook. Columbia basin states.
1968. Climatological
Tol. III,
UI, Part
Part B. 641 p.
Hourly data. Vol.
Naval Weather Service Command. Environmental Detachment. 1971.
synoptic meteorological
Summary of :synoptic
meteorological observations, ocean weather
U?
HNht and
and "P" Asheville, North Carolina. 316 numb.
stations
stationsHNht
leaves.
1910.
10. Moored spar
Neshyba, Steve, D. A. Young and J. H. Nath. 19
buoys for ocean research. Oceanology
Oceanology International,
International, April
1970, p. 22-26.
Skarr, John. 1955. On the measurement of precipitation at sea.
1 -32.
Geofysiske Publikasjoner
Publikasjoner99(6):
p. 1-32.
Geofysiske
(6): p.
Supan, A. 1898. Die j.hr1ichen Niederschlagsmengen auf den
Meeren.
Meeren. -Petermanns
Petermanns Geographische Mitteilungen
Mitteilungen 44:179-182.
(Cited in: Jacobs, W. C. 1968. The seasonal apportionment
climotology,
of precipitation over the oceans. In: Eclectic clirnotology,
ed. by Arnold Court. Corvallis, Oregon State University
65)
Press. p. 65)
Precipitation over the North Atlantic Ocean.
Quarterly Journal of the Royal Meteorological Society 87:147-
Tucker,
0. B.
Tucker, G.
1961.
158.
U. S. Department of Commerce. Coast and Geodetic Survey. 1951.
United States coast pilot, Pacific coast, California, Oregon,
7th ed. Washington, D.C. 578 p.
and Washington. 7th
74
U. S. Department of Commerce. Environmental
Environmental Science Services
Administration. Environmental Data Service. 1970. Daily
22, 1970. n.p.
5-March22,
weather maps. Jan. 5-March
Department of
of Commerce.
Commerce. National
National Climatic Center. 1970.
U.. S.: Department
144.
Reference manual. WBAN hourlysurface
hourlysurface observations
observations 144.
Rev. ed. Asheville, North Carolina. 14 p.
U. S. Department of Commerce. National Ocean Survey. 1911.
C. & C. S. Chart 5021. Monterey Bay to Coos Bay. 5th Ed.
Washington, D.C. n.p.
1971.
C. & C.
0. S.
S. Chart 5022. Cape Blanco to
Cape Flattery. 5th ed. Washington, D. C. n. p.
U. S. Department of Commerce. National Oceanic and Atmospheric
Administration. Environmental Data Service. 1970-1971.
Climatological data. Oregon. Vol. 76-77.
1970-1971. Hourly precipitation data. Oregon.
Vol. 20-21.
1971-1972. Climatological data.
Oregon. Vol.
77-78.
U. S. Department of Commerce. Weather Bureau. 1953-1966.
Vol. 4-17.
National summary.
summary. Vol.
Climatological data. National
1955-1965. Climatological data. Washington,
annual summary. Vol. 59-69.
World Meteorological Organization. 1962. Pre cipitation measure ments at sea. 18 p. (Technical Note no. 47).
Oberflachensalzgehalt, Verdunstung und
Wust, Georg.
Ceorg. 1936. Oberfl.chensa1zgehalt,
Lánderkundliche Forschung
Forschung
Niederschlag
aufdem
demWeltmeere.
Weitmeere. Lnderkund1iche
Niederschiag auf
Jacobs,
(Cited
in:
in:
Jacobs,
(Cited
9.
(Festchrift .Norbert
.Norbert Krebs),
Krebs), pp.
pp. 347-35
347-359.
exchange between
betweenthe
thesea
seaand
andatmosatmosW. C. 1951 The energy exchange
Bulletin
of
the
Scripps
phere and some of its consequences.
Institution of Oceanography, University of California, 6: p. 28).